The act of doing isn’t the tough part, AI’s quality of execution gets better day by day.
Judgement, ability to select what to and what not to do, enhance our ability to orchestrate these experts we have.
Judgement is an accumulation of experiments.
Experiment carried to close the inefficient gaps in our work. If it succeed, we’re rewarded with extra time. If it fail, we iterate.
This cycle repeated will create a new pattern.
New side project: network graph linked by post backlinks. Comparison between Google Stitch and Claude Code with frontend design skill.
- Experiments are free context to test new tools.
- Experiments accumulates judgement on how the new tool differs.
Claude Code is still great for quick vibe coding. It’s an all-in-one architect + construction + internal design.
Google Stitch is an amazing internal designer. The canva visualize the specification and branches, it is valuable after we experience the whole cycle. We know which stage the new tool brings value.